the geography of venture capital contractsleeds-faculty.colorado.edu/bhagat/geographyventure... ·...
TRANSCRIPT
Electronic copy available at: http://ssrn.com/abstract=1361827
The Geography of Venture Capital Contracts
by
Ola Bengtsson* and S. Abraham Ravid**
First draft: June 2008
This draft: March 2009
Abstract
We show that geographical elements and regional culture can play an essential role in contract design in addition to the influence of more “traditional” determinants such as information and agency problems or the nature of legal institutions. Across 1,800 financial contracts written between U.S. entrepreneurial companies and U.S. Venture Capital (VC) investors, we show that contracts include significantly fewer investor-friendly cash flow contingencies if the company is located in California or if the lead VC is more exposed to the California market. The regional differences in contract design can, to some degree, be explained by the level of concentration of local VC markets. We also show that when the geographical distance between a VC and a company is greater, contracts give high-powered incentives to entrepreneurs by including more investor-friendly cash flow contingencies. This latter finding supports the view that geographical proximity enhances monitoring and soft information. However, the “California effect” persists even after we control for distance and VC market concentration.
JEL Classification: G24, L26 Keywords: Venture Capital, Entrepreneurship, Financial Contracting, Geography *Cornell University and **Rutgers University and the Wharton School, University of Pennsylvania. Contact information: [email protected]; [email protected] or [email protected]. We are grateful to VCExperts and Joseph Bartlett for access to the contract data, to a number of attorneys and VC partners for helpful comments, and to Sonali Das, Emily Feng, Bonny Lee, Cathy Chenxi You, Brian Steinmetz and Vikas Patel for valuable research assistance. All remaining errors are our own.
Electronic copy available at: http://ssrn.com/abstract=1361827
1. Introduction and Literature Review
A large body of literature on financial contracts and security design examines how the
allocation of cash flow contingencies and control rights is related to the characteristics of the
contracting parties, the institutional environment, and the quality of the legal system.1 This paper
suggests that contract design may depend not only on such “traditional” ingredients but also on
geographical distance between the investor and the borrower and, importantly, on their specific
locations. The distance effect can be motivated by lower monitoring costs and the presence of soft
information. However, our finding that contracts depend on geographical location is hard to
reconcile with most existing theoretical models of contract design, since we study contracts from
one country exclusively (the United States). The differences that we document therefore cannot
be attributed to variations in tax or bankruptcy codes, or to differences in securities laws or legal
enforcement costs. Instead, the finding that location matters suggests that regional culture and
local customs may play a role in contract design.
We study 1,800 venture capital (VC) contracts drawn between entrepreneurial firms and
their VC investors. The VC industry is interesting in its own right given its importance to the U.S.
economy, but it also represents a useful empirical testing ground for contract theories (Kaplan &
Stromberg, 2003; Hart, 2001).2 VCs are sophisticated investors, well-versed in incentive
contracts, who provide financing to young, high-growth companies that typically suffer from
severe agency and information problems. The contracts that VCs receive in exchange for their
investments are complex and non-standardized, and have been shown to share many of the
1 There are too many papers to be listed here; however, in the context of finance they can include Townsend (1979), Allen & Gale (1988), Harris & Raviv (1989, 1995), Madan & Soubra (1991), Boot & Thakor (1993), Fluck (1998), Zender (1991), and, in the specific context of venture capital or startup firms, also Admati & Pfleiderer (1994) and Ravid & Spiegel (1997). 2 Some basic statistics illustrate the economic importance of the VC industry: Annual VC investments in 2007 reached $30.7 billion, 344 venture-backed companies went public from 2002–2007, and venture-backed companies provided 10.4 million jobs and $2.3 in revenue in 2006. Many of today’s high-profile companies were once backed by VCs, including Microsoft, Amgen, Google, Facebook, and FedEx.
Electronic copy available at: http://ssrn.com/abstract=1361827
features predicted by contract theory (Sahlman, 1990; Gompers, 1988; Kaplan & Stromberg
2003, 2004; Bengtsson & Sensoy, 2008; Cumming, 2008).
The VC industry is also a promising testing ground for our purposes because it is
probably the largest and most developed capital market in which geographical and cultural factors
can play an essential role. Unlike public debt and equity markets, the U.S. VC market is not
nationally integrated but exhibits a considerable degree of fragmentation. VCs often invest locally
(Gupta & Sapienza, 1992; Norton & Tenenbaum, 1993; Stuart & Sorensen, 2001; Bengtsson,
2008) and also form strong syndication networks with other local VCs (Hochberg, Ljungqvist,
and Lu, 2008). Geographical and cultural factors may arise from the presence of formal and
informal networks between venture-backed companies. Gompers, Lerner and Scharfstein (2005)
show that many new venture-backed companies are spawned from local public companies that
were once venture-backed. Lindsey (2008) presents evidence that strategic alliances between
venture-backed companies are commonplace, especially for companies that share a VC investor.
The fact that many venture investors were themselves previously active as entrepreneurs
(Zarutskie, 2008) adds another possible channel through which cultural and geographical as well
as other informal factors can affect the VC industry.
The most pronounced geographical segmentation of the U.S. VC market is the difference
between the “West Coast” and the “East Coast. Saxenian (1996) suggests that differences in
corporate and investor culture drive the vastly different fortunes of Silicon Valley and the Route
128 corridor in Massachusetts. Both regions attracted major high-tech companies at the start of
the recent computer age. In fact, the industry structure and high-technology employment in the
two regions were roughly similar in the mid 70s. Yet, since the 1990s, Silicon Valley has been
vastly more successful (figure 1, p. 3, ibid). In Saxenian’s view this is attributable to looser
boundaries between firms and the informal office culture (T-shirts) and networks in Silicon
Valley versus a more formal culture (dress shirts) that is less conducive to change and progress in
the Route 128 area.3
Our empirical work shows that the less formal culture in California, and in particular in
Silicon Valley, is associated with a less investor-friendly allocation of cash flow contingencies in
VC contracts. This difference in contract design is found if either the company or the lead VC,
who acts as the primary investor in the investment round, is located in this region. An analysis of
control rights provides further evidence that location matters—contracts used by Silicon Valley
lead VCs are less investor-friendly since they allocate fewer board seats and protective covenants
to investors. Importantly, these results are unlikely to reflect differences in company or
entrepreneur quality, but are instead likely to be motivated by regional differences in contracting
style. In interviews with executives at VC firms as well as lawyers representing such firms, we
have been repeatedly told that contracts on the West Coast are less harsh towards entrepreneurs
because VCs there take a more partner-like approach to investing as opposed to the banker-like
approach that is common in other regions, in particular on the East Coast.
We show further that contract design in the VC industry carries over between regional
markets, with contracts including fewer investor-friendly cash flow contingencies if a non-
California lead VC has had considerable exposure to investments in California. This result is
consistent with VC’s acquiring knowledge about how to structure contracts optimally when they
invest in one market and then applying this knowledge to other markets. A similar result is found
by Kaplan, Martel, and Stromberg (2007), who show that non-U.S. VCs learn how to structure
contracts from co-investing with U.S. VCs.
The results pertaining to location effects hold after we account for all observable
differences with regards to round, company, VC, and founder characteristics and even variables
that capture the concentration of the regional VC market. We use distinct proxies for the
3 A theoretical model that endogenizes regional differences in culture is provided by Landier (2006). His model demonstrates how differences in investors’ evaluation of entrepreneurial failure could arise as multiple equilibria in a fully rational setting.
concentration of VCs and venture-backed companies and show that more highly concentrated
markets feature contracts that include fewer investor-friendly cash flow contingencies. This result
is similar to the finding of Degryse and Ongena (2005) that bank interest rates are lower when a
borrowing firm has access to a greater number of competing lenders that are located nearby. In
our setting, competition between VCs allows entrepreneurs to negotiate contracts that are less
investor-friendly in their design.
Finally, we show also that distance between the lead VC and the funded company matters
in contract design, as predicted by several theories addressing soft information and monitoring
costs. Contracts involving VCs and entrepreneurs allocate more investor-friendly cash flow
contingencies and thereby provide for more high-powered payoff incentives to the entrepreneur
the greater is the distance between the firm and the lead VC investor. Interestingly, however, the
“California effect” remains even after we account for the distance and concentration effects.
This paper contributes to several streams in the literature. We add to the small number of
studies that explore geographical impact on VC contracts. Unlike studies of international
differences in VC contracts (Lerner & Schoar, 2005; Kaplan, Martel & Stromberg, 2007; Bottazi,
DaRin, & Hellmann, 2008) and VC investment decisions (Cumming et al, 2008), our paper is the
first to study the role of geography within a country. As noted, this means that our results cannot
be explained by differences in the legal system, rule-of-law, accounting transparency, bankruptcy
procedures, taxation, etc.4 Interviews with lawyers and legal scholars have confirmed that there
are no institutional reasons why U.S. VC contracts should vary by company or VC location.5
Our findings also have important implications for the empirical testing of models of VC
contract design to explain why various types of convertible securities are used in VC investments
(see Berglof, 1994; Hellman; 1998, 2006; Cornelli & Yosha, 2002; Casamatta, 2003; Schmidt,
4 Gilson & Schizer (2002) discuss how the prevalent use of convertible preferred equity in VC investments could be motivated to some degree by the U.S. tax code. 5 The only potentially relevant between-state institutional difference is in the enforcement of non-compete employment agreements. As discussed in Section 3, this cannot explain the difference we have found between contracts in California and Silicon Valley.
2003; Repullo & Suarez, 2004). Because geographical factors have real-world relevance for how
VC contracts are designed, they should be included as controls in any empirical analysis of cash
flow and control rights, and when distance is studied it is important to separate Silicon Valley and
California from other locations
We also contribute to the growing body of finance literature on geography and home
bias, issues that in recent years has received increasing attention. Grinblatt and Keloharju (2001)
find that portfolios of retail investors are biased towards local companies. Huberman (2001) finds
that this higher fraction of local stocks in investor portfolios is due primarily to familiarity with
these stocks. In contrast, Ivkovich and Weisbennar (2005) show that retail investors are better
informed about local investments and these local investments are associated with higher returns.
Coval and Moskowitz (2001) document a similar local bias in the portfolios of mutual fund
investors and also show that geographically proximate institutions have information advantages.
If both retail and institutional investors bias their portfolios towards local stocks, then a large
fraction of the trading volume is likely to originate locally. Kedia and Zhou (2007) show that a
large presence of local market makers significantly reduces both quoted as well as effective
spreads. Similarly, Malloy (2005) documents how geographically proximate analysts make fewer
forecast errors and Uysal, Kedia, and Panchapagesan (2008) show that local acquirers enjoy
higher returns in mergers and acquisitions. Schultz (2003) shows that geography provides an
information advantage in the context of an IPO syndicate.
The common theme among all these papers is that business-related activities and social
interactions (golf games, Rotary club meetings, etc.) between executives may provide each side
with better information and a more favorable view of one another. Local media are also more
likely to pay attention to local companies and thus make information easily available to local
market actors. For active investors such as VCs, home bias is particularly pronounced because
geographical proximity could lower pre-investment screening costs as well as post-investment
monitoring costs. All venture-backed companies have VCs represented on their boards of
directors and VCs frequently visit their portfolio companies to interact with the founders and
management (Gorman & Sahlman, 1989). Lerner (1995) finds evidence consistent with the notion
that VC oversight of private firms is related to geographical distance—VCs that are
headquartered close to a portfolio company are significantly more likely to take a seat on the
board of directors. We add to this niche in the literature by looking in depth into a large sample of
contracts, and investigating the distance, concentration, and cultural components of contracts.
Finally, our paper contributes to the small body of literature that attempts to empirically
test the validity of different contract theories and provide real-world evidence on contract design.
In addition to being the subject of VC studies, contract theory has been tested in two other broad
areas, namely, biotechnology and movie studies. Biotechnology papers focus on the distribution
of various rights between contracting firms (see, for example, Lerner & Merges, 1998). The film
industry is characterized by interesting and complex contracting. There is generally less data
available on film industry contract design than on VC or biotechnology contracts, but outcomes
are much better known. Chisholm (1997) analyzes several dozen actor contracts and shows that
more experienced actors are more likely to receive a share contract, supporting some lifecycle
compensation theories. Palia et al. (2008) focus on co-financing agreements and test theories of
the boundaries of the firm, whereas Goetzmann et al. (2008) discuss screenplay sales contracts,
focusing on soft information.
In other industries there is sparse empirical work on contract design, due to the scarcity of
data. Banerjee and Duflo (2000), for example, show that better reputation (in Indian software
companies) leads to lower prevalence of fixed-payment contracts, which provides greater
incentives to firms than “contingent” contracts. They discuss software projects, and the
“contingent” contract is essentially a time-and-materials contract, that is, a contract with no
specific price estimate. While each industry is characterized by different institutions, most studies
support some of the major features predicted by the theory. In addition to our distance and
location variables, we therefore include in our tests all contractual variables previously studied in
connection with VC contracts, which are closest to us in terms of methodology. In particular, we
include variables used in Kaplan and Stromberg (2003) and Bengtsson and Sensoy (2008).
The rest of the paper is organized as follows—the next section describes our data and the
coding of VC contract terms. The third section explores regional differences in contract design,
and the fourth section discusses the role of VC market concentration on contract design. The fifth
section includes tests of the role of distance in contract design. The last section summarizes our
finding and concludes.
2. The Data
Sample
We study a sample of contracts between U.S. early-stage private companies and their VC
investors. An overview of the sample is presented in table 1. The contract data are collected and
coded with the help of VCExperts, and cover 1,800 investment rounds in almost 1,500 unique
companies (this method of classifying VC contracts is common in the literature; see, for example,
Kaplan & Stromberg, 2003). Our sample is the largest dataset of VC contracts studied by
academic researchers to date, and includes more than ten times more companies than the samples
used by Kaplan and Stromberg (2003, 2004) and Cumming (2008). Our deals are recent, with
83% of investment rounds being closed in 2006 and 2007. The majority of companies are from
high technology or life science industries.
A substantial fraction of the contracts in our sample (90%) represents investment rounds
in which more than one VC firm participated. In such syndicated rounds, one investor typically
takes a lead role and conducts most of the pre-investment screening and post-investment
monitoring. The lead VC also drafts the proposed contract and takes point during contract
negotiations with the entrepreneur. Because contract design is influenced primarily by the lead
VC, we conduct our empirical analysis at the contract level using information about the lead VC.
We identify the lead VC as the investor taking the largest stake in the round. In cases where two
or more VCs take the same stake, we define the most experienced VC, measured by historical
number of portfolio companies, as the lead investor. Our sample is restricted to contracts for
which the lead VC was headquartered in the U.S.
We use zip-code data to measure the exact location of lead VCs and companies in our
sample. The data exhibit, as expected from a study sampling U.S. VCs, a strong “California”
element—California houses about 35% of the sample companies and 35% of the VCs that were
lead investors in the round. In California, the Silicon Valley takes the largest single cluster with
about 13% of companies and 25% of lead VCs. Many famous VCs, including New Enterprise
Associates, Sequoia Capital, U.S. Venture Partners, and Kleiner, Perkins, Caufield & Byers, are
headquartered along Sand Hill Road in Silicon Valley. The second largest cluster is from
Massachusetts, with 16% of all companies and 19% of all lead VCs, many of whom are located
along Route 128. Other large VC markets include Texas (especially Austin) and North Carolina
(especially the Research Triangle of Raleigh, Durham, and Chapel Hill).
Summary statistics are reported in table 2A. Consistent with the results of earlier studies
(Gupta & Sapienza, 1992; Norton & Tenenbaum, 1993; Stuart & Sorensen, 2001; Bengtsson,
2008), we find that VCs prefer to invest in companies that are located close to their headquarters.
One in five companies is located no more than 10 miles from their lead VC and 42% of
companies are located no more than 50 miles away.6
We match each contract with an investment round in VentureEconomics and obtain
variables that measure company and lead VC characteristics. We also hand-collect data on the
characteristics of the founding team. For about half of our sample, we obtain data from
VCExperts and VentureEconomics on the pre-money valuation of the company. This valuation
number is a negotiated term that captures the VC’s assessment of the company before the
financing round is closed. The average sample company raised $11 million dollars at a pre-money
6 One rule of thumb in VC investing is the so-called “20 minute rule,” according to which a VC should be no further away than a 20-minute drive from a portfolio company. Our data shows that this rule is not always obeyed.
valuation was $48 million. For a subset of our sample we also have data on the contractual
allocation of board seats and protective covenants that give VCs certain veto rights over
important business decisions. We use these data in the analysis later.
Contract Terms and Contract Harshness
Each of the 1,800 unique contracts is coded along six important contractual dimensions,
namely, cumulative dividends, liquidation preference, participation, anti-dilution rights,
redemption, and pay-to-play. The six contract terms jointly define the cash flow contingencies
that are attached to the preferred stock that VCs receive in exchange for their investments.7 In
other words, the contract terms determine how many additional cash flow contingencies are given
to the holder of one share of preferred stock. As shown by Kaplan and Stromberg (2003) and
Bengtsson and Sensoy (2008), most terms that are included in VC contracts are favorable to the
VC and especially favorable if company performance is poor.8
The exact meaning and economic importance of each cash flow term that we study is
described below. Table 2A provides an overview of the contract terms and reports their frequency
in our sample. We code each contract term as 0 or 1 based on how favorable it is to the VC,
where a value of 1 means that the contract is “harshest” for the entrepreneur and the other
existing owners of the company, or alternatively more favorable for the VC who invests in a
round. This coding methodology is similar to that used by Bengtsson and Sensoy (2008), with
main difference being that they code some contract terms as 0, 1 and 2. All empirical results
presented in this paper are qualitatively the same if we use the coding approach of Bengtsson and
Sensoy (2008).
7 Redemption rights, which permit VCs to sell their shares back to the company after a pre-determined time period, is a cash flow contingency in the sense that it represents a valuable put option issued by the company to the VC. 8 The exception is pay-to-play, which when included does not favor the VC. We code pay-to-play inversely.
While the six contract terms we study are functionally similar, they could be included or
excluded in the contract independently of each other. We aggregate the six binary variables to an
index labeled Aggregate Contract Harshness (ACH). ACH takes values from 0–6, where 0 is a
contract that includes a minimum of investor-friendly cash flow contingencies and 6 is a contract
that includes all possible investor-friendly cash flow contingencies. As reported in table 1, the
average ACH value is 2.59 and the median is 3. Since we are interested in the joint contractual
allocation of cash flow contingencies, our primary variable of study is ACH. We also study each
cash flow contingency in separate empirical tests.
Appendix A provides a detailed description of the meaning of the six contract terms, their
financial implications and exactly how they enter into our index ACH. An abbreviated description
of this is provided in table 2B.
3. Contract Terms and Location of Company and Lead VC
We now proceed to our analysis of geography and contract terms. We first study the
relationship between contracts and geographical location, then proceed to explore the relationship
between contracts and market concentration, and finally study the role of distance in contract
design.
Aggregate Contract Harshness
Table 3A provides the first data classification suggesting a strong geography component.
In panel A we present univariate comparisons showing that both lead VC and company location
matters for contract design. VCs in California tend to offer fewer investor-friendly terms, and
companies based in California also tend to receive fewer investor-friendly terms. The effects are
even stronger if either the lead VC or the company is located in the Silicon Valley. Kaplan and
Stromberg (2003) also find that contract terms are different for a California location of the
company. In their case, California contracts use less-explicit performance benchmarks, have
lower claims for the VC, and include fewer redemption rights, results that are consistent with our
findings. We add to the geography result of Kaplan and Stromberg by showing that, in addition to
a California effect for company location, there exists an important California effect for the
location of the lead VC. Also, we show that the California effect is particularly pronounced for a
Silicon Valley location.
The geographical impact on contract design is economically large—a contract between a
company and lead VC that are both located in the Silicon Valley is almost one ACH unit less
investor-friendly than a contract between a company and lead VC that are both located outside
Silicon Valley. This regional difference in contract design represents about one standard
deviation of the cross-sectional distribution of ACH, and is notably larger than differences based
on a sort on empirical proxies for agency and information problems (and therefore conceptually
should matter for contract design). As shown in table 3A panel C, a contract offered to a company
that has a serial successful founder, has secured a high round amount, and is financed by an
experienced lead VC has only 0.7 unit of ACH less than a company that has no serial successful
founder, has secured a low round amount, and is financed by an inexperienced lead VC.10
Figure 1 provides a coarse illustration of the pattern of regional differences in contract
design. We calculate the aggregate contract harshness (ACH) of the average contract for each
U.S. state. A darker color represents contracts that are more investor friendly, and we note that
VC contracts include fewer investor-favorable cash flow contingencies on the West Coast than on
the East Coast.
Table 4 is a first multivariate exploration of the harshness of contract design and it
focuses on the California effect. We run an ordered logit regression with ACH as the dependent
variable and include all commonly used contract-theoretical variables as well as variables relating 10 See Bengtsson and Sensoy (2008) for an empirical motivation of these variables.
to the location of the company and the lead VC. All regressions control for the company industry
(as per VentureEconomics 10-level classification) and year of the sample round. We include
company age, number of previous financing rounds, and the dollar amount raised in the sample
round. We also include variables that capture whether any of the company’s founders has
previously started a venture-backed company and whether this company was successfully taken
public or acquired by another corporation. These company, round, and founder characteristics
capture, in different ways, the overall quality of the company and the magnitude of the agency
problem that its lead VC investor faces. Kaplan and Stromberg (2003, 2004) show that these
factors affect the design of real-world VC contracts. Because VC contracts are also affected by
VC characteristics (Bengtsson & Sensoy, 2008), we include the number of VCs, and the lead
VC’s investment experience and type (i.e., private partnership VC, corporate VC, financial VC,
etc).
The multivariate analysis confirms the results of the univariate comparison, showing a
strong California effect on contract design. This effect seems to be largely a Silicon Valley
effect—in other words, among California companies, a Silicon Valley location provides an extra
boost to the leniency of the contract. As shown in regression model 8, contracts become more
investor-friendly as the lead VC is located farther from the Silicon Valley.
We also note that some important control variables have significant coefficients that are
directionally consistent with results from previous studies. Consistently with Kaplan and
Stromberg (2003, 2004), for example, we find that companies with greater maturity include more
investor-friendly cash flow contingencies. We also validate the findings of Bengtsson and Sensoy
(2008), whose sample to a large extent overlaps with ours, that companies raising larger amounts
of VC financing from more-experienced VCs have fewer investor-friendly cash flow
contingencies.
With the goal of further examining the extent of the California effect, we show that
investor-friendly cash flow contingencies are used less frequently not only by California VC
firms but also by non-California VCs that have greater exposure to the California market.
Restricting the sample to lead VCs headquartered outside California, we use previous investment
experience to create two new explanatory variables. The first variable is the lead VC’s California
investment experience, which measures how many times the VC has previously invested in
companies located in California. The second variable is the lead VC’s California syndication
experience, which measures how many times the VC has previously invested in a round that was
syndicated with a VC headquartered in California. We find that any California connection
significantly improves contract terms for the entrepreneur. This is perhaps the most convincing
piece of evidence favoring an explanation in terms of a different “contracting style” in California
as described by Kaplan and Stromberg (2003, p.299). The result that contract design carries over
between markets is also consistent with the argument in Kaplan, Martel, and Stromberg (2007)
that VCs learn about a certain contracting style by co-investing with other VCs who use and
understand that contracting style.
Regional Culture and Customs Explanation
Conversations with VCs and attorneys specializing in VC contracts that were intended to
gauge the source of the California effect seem to point to a geographical dispersion of opinions
that is not tied to specific legal or tax provisions. Quotes from two reputable VC attorneys
illustrate the industry perception that there are important regional differences in contract design.
Eduardo C. LeFevre (of Foley & Lardner LLP) says: “There is also a growing awareness of the
differences between ‘East Coast’ and ‘West Coast’ financings, primarily with respect to regional
differences in valuation, liquidation preference, and number of later stage financings.” Alan
Bickerstaff (of Andrews Kurth LLP) adds: “The terms of VC financings are fairly customary,
with nuances unique to each deal and geographic region. For example, East Coast VCs tend to
require founders personally to make certain representations and warranties whereas this practice
is virtually nonexistent in West Coast deals” (Falls, 2008, p. 90 and p. 101). In fact, a VC attorney
told us that, when the National Venture Capital Association tried to come up with a common
template for VC contract provisions, “Western” VCs thought that what “Eastern” VCs were
proposing was way too harsh. This also agrees with the thrust of Saxenian’s (1996) argument.
Thus, our results are consistent with explanations of geographical differences in contract design
that refer to regional culture and customs.
Alternative Explanations
A possible alternative explanation of our results is that the observed differences in
contract design reflect not regional differences per se but differences in company quality.
Previous studies of VC contracts show that higher quality venture-backed companies, which
present investors with lower risks, are able to negotiate contracts with fewer investor-friendly
cash flow contingencies (Kaplan & Stromberg, 2003, 2004; Bengtsson & Sensoy, 2008). If such
companies are systematically more likely to be located in California, then it is possible that
regional differences in company characteristics explain the California effect in contract design.
Although it is hard to rule out selection in the VC setting econometrically (because of data
limitations and the absence of good instruments), we have two reasons to believe that our results
cannot be fully explained by company quality.
First, all our regressions control for company age, round amount, and the founders’ track
record, all which are likely to be relatively good proxies for company quality. The results also
hold after controlling for VC experience, which due to selection in the VC market also is a proxy
for company quality (Sorensen, 2008).
Second, the magnitude of the California effect is, as discussed, so large that the
differences in company quality must be substantial in order to fully explain it. In unreported
empirical tests, we test whether successful outcomes are more common for companies located in
California. To do these tests, we use VentureEconomics to extract data on all U.S. venture-backed
companies that were initially funded between 1983 and 2002.12 We find that neither IPOs nor
acquisitions are more common for companies located in California than for those located in other
U.S. states, after controlling for the control variables used in our analysis of contracts. Given that
companies in California historically have not been more successful, it is in our view highly
unlikely that company quality can explain the significantly less investor-friendly contract design
associated with California.
Before we proceed with the analysis, it is also important to emphasize that, because all
companies and lead VCs that we study are located in the U.S., our results cannot be explained by
differences in tax code, bankruptcy procedures, legal infrastructure or enforcement of financial
contracts.13 Interviews with legal scholars and practicing VC lawyers confirm the view that no
institutional factor suggests that the design of VC contracts should vary between US states. To the
best of our knowledge, the only potentially relevant institutional difference between U.S. states is
the ability to enforce non-compete clauses in employment contracts. Such contracts are notably
more difficult to enforce in California courts. This difference is, however, very unlikely to
explain our results since we observe important differences in contract design between Silicon
Valley and other locations in California, for which state laws are identical. Also, between-state
differences in the enforcement of non-compete clauses cannot explain why, after controlling for
company location, we observe differences based on VC location and VC exposure to the
California market.
12 We exclude companies funded after 2002 because venture-backed companies typically need between four and six years from initial funding to successful exit. 13 In untabulated regressions we have also controlled for the state in which the company is legally incorporated (which is most commonly Delaware, followed by California). The irrelevance of incorporation for contract design is illustrated by our findings that all reported results remain unchanged and the estimated coefficients on incorporation-state dummies are not significant.
Separate Contract Terms
The next step of the analysis of location and contract design is to study each cash flow
contingency separately. Table 3B panel A reports the results for comparisons based on company
and VC location. The analysis shows that individual contract terms are overall less investor-
friendly in Silicon Valley. The notable exception is pay-to-play, which is more common if the
lead VC or company is located in Silicon Valley (pay-to-play is not VC favorable and coded as 1
if it is not present). Thus, while the average Silicon Valley contract includes fewer investor-
friendly cash flow contingencies, the lower likelihood of a pay-to-play provision implies that such
contingencies are not void if VCs choose not to invest in a follow-up financing round.
The most pronounced difference between the terms of Silicon Valley contracts and those
of other contracts is in the prevalence of cumulative dividends and redemption rights. The VC
attorney David K. Levine (of Snell & Wilmer LLP) confirms this specific finding: “It may be a
bit more common for VCs based on the East Coast to require dividends that accrue (or cumulate)
but such cumulative dividends provisions are quite rare in West Coast based deals” (Falls, 2008,
p. 129).
Table 5A adds probit regressions in which each separate contract term in turn is the
dependent variable. In addition to “VC in Silicon Valley” and “Company in Silicon Valley,” our
independent variables include the full set of contract-theoretical control variables. Interestingly,
as shown in regression models 7–11, Silicon Valley is relatively similar to other geographical
areas when we compare other important deal dimensions such as round amount, number of VCs
in the round and valuation. This suggests again a difference in culture and style rather than in
tangible legal premises. We find, however, that companies headquartered in Silicon Valley tend
to give VCs as a group a larger ownership stake in a round, which is suggestive evidence that
investors, at least to some degree, compensate for the use of contracts with fewer investor-
friendly cash flow contingencies by demanding higher ownership stakes.
Control Rights
Finally, we analyze whether regional differences in the use of cash flow contingencies
also extend to the allocation of control rights between VCs and entrepreneurs. As noted, for a
subset of our sample we also have data on the contractual allocation of board seats and negative
covenants, which give VCs collective veto rights over important business decisions. Table 5B
presents regressions similar to the specifications in table 5A but with different measures of a
number of control rights as dependent variables.15 We include dummies that capture whether the
company or VC was headquartered in Silicon Valley.
Our analysis of control rights demonstrates that VCs receive fewer board seats (model 1)
when the lead VC is headquartered in Silicon Valley, and are thereby less likely to have a board
majority (model 3). VCs headquartered in Silicon Valley furthermore use contracts with fewer
covenants (model 4) such as the right to block the company from making changes to its business
model (model 7), take on new debt (model 8), incur capital expenditure (model 9), enter into a
joint venture or strategic alliance (model 10) or initiate a recapitalization or reorganization (model
11).16 These results on control rights are important because they demonstrate that VCs
headquartered in Silicon Valley do not agree to fewer investor-friendly cash flow contingencies
in order to compensate for more investor-friendly control rights. This is further evidence that
contract design reflects regional differences in style and culture, with Silicon Valley investors
using contracts that are overall less harsh towards entrepreneurs. In untabulated regressions we
replace the Silicon Valley dummies with California dummies and obtain qualitatively similar
results.
15 Note that sample size varies across specifications in this table because of incomplete data on board membership. 16 The total number of covenants used as a dependent variable in model 4 includes a total of 18 protective provisions. Debt and CapEx covenants typically specify a dollar amount above which the covenant is binding.
4. Contract Terms and VC Market Concentration
Our results thus far have demonstrated a significant regional effect in VC contracts,
which, again, cannot be attributed to legal and institutional differences, as all contracts are signed
by U.S. companies and U.S. lead VCs. It may be, however, that California is unique not because
of cultural factors or regional customs, but rather because this market has the highest
concentration of VCs and venture-backed companies. We now proceed to analyze this
explanation more formally.
We create a variable that measures the number of active VCs in the state where the
company is located. Figure 2 illustrates the number of active VCs in each state, where a darker
area represents a higher concentration. Figure 1 illustrates the aggregate contract harshness
(ACH) of the average contract, with a darker area representing a more investor-friendly contract.
A comparison of figures 1 and 2 clearly illustrates a inverse relationship between ACH and the
number of active VCs in a state. Hence, VC contracts include on average more investor-friendly
cash flow contingencies in U.S. states that have less-concentrated VC markets.
We confirm this idea in multivariate regressions shown in table 6. We regress ACH on
the company, lead VC, and round variables and also include a measure of VC concentration. VC
concentration is positively correlated with ACH, whether or not it is measured by the number of
active VCs in a state, the number of active VCs in a region (using the Census 9-region
classification of the U.S. states), the number of venture-backed companies in a state-industry
segment or the total dollar amount raised by venture-backed companies in a state-industry
segment.
The result holds even after we control for whether the company or lead VC was located
in California (models 4-8). Importantly, the coefficients on the California dummies remain
negative and significant. Thus, companies that are located in California include fewer investor-
friendly contract terms partly because there are more active VCs or more VC funding in this state,
but other regional or cultural differences still seem to affect contract design.17
One plausible explanation for the empirical association between contract design and VC
market concentration is that a more active VC market increases an entrepreneur’s bargaining
power in contract negotiations. Degryse and Ongena (2005) show that bank interest rates are
lower when a borrowing firm has access to more competing lenders that are located nearby,
suggesting a role for competition in contract design. If a VC offers a contract with an onerous
number of investor-friendly cash flow contingencies then the entrepreneur can reject, or threaten
to reject, the financing offer and seek funding from another VC.
5. Contract Terms and Distance between Company and Lead VC
Our final set of tests considers another aspect of the location effect on contract design,
namely, whether the relative distance between company and lead VC also influences how
contracts are written. Papers on soft information (see Stein, 2002; Petersen & Rajan, 2002; Berger
et al. 2005; Petersen, 2004; and Uzzi, 1999) suggest that, in the presence of soft information and
monitoring costs, smaller local banks may be better suited to serve local customers. In our setting,
if the VC and the entrepreneur are on close personal terms, they may need only the proverbial
handshake rather than a complicated contract with harsh cash flow contingencies. The evidence in
Lerner (1995) is consistent with the idea that geographical distance affects how a VC interacts
with companies in its portfolio.
We first use a zip code database to look up the longitude and latitude of the main office
for each sample company and lead VC, and then calculate distance in miles using the Haversine
formula, which takes into account the curvature of the Earth. Some evidence suggesting that
distance matters is found in the univariate comparisons shown in table 3A panel B. Companies
17 In untabulated regressions, we include square measures of our variables that capture VC and company concentration. The coefficients on the California dummies remain significant after controlling for such potential non-linearity between ACH and concentration of the VC market.
that are located geographically closer to their lead VCs are significantly less likely to include
investor-friendly contract terms. Such contracts translate into more high-powered payoff
incentives for entrepreneurs because the cash flow contingencies that we study increase the VC’s
payoff more in bad states of the world than in good states of the world.
As an illustration of our results, the average ACH is 2.46 when company and lead VC are
located in the same state, as compared with 2.72 when the company and lead VC are located in
different states. For a company outside California, a contract from a within-state lead VC has an
ACH of 2.87 whereas a contract from an out-of-state non-California lead VC has an ACH of
3.03. However, this company would get an average ACH of 2.44 from a California lead VC. In
other words, contracts are less investor-friendly when the company and lead VC are located close
to one another, except that contracts are always less investor-friendly if the lead VC is
headquartered in California. Thus, the data exhibits both a distance effect and a California effect
but the latter seems to be economically more important.
Table 7 confirms the distance results in a multivariate setting. Regression models 1–5
include sample companies located in California and models 6–10 include companies located in
other states. The regressions are similar to those presented in table 4 and include all controls used
previously, but for space considerations we show only the geography and California effects. The
California effect is shown to be as significant here as it is in table 4.18 However, distance seems to
be important as well.
Finally, we return to table 3B panel B to explore the relationship between distance and
individual contract terms. The effect of distance holds for all contract terms except liquidation
preference and pay-to-play. Companies located in California are less likely to sign contracts with
investor-friendly cumulative dividends, anti-dilution, and redemption rights if they receive
financing from a California lead VC. For companies located in other states, cumulative dividends
18 All previously reported regression results related to VC and company location are qualitatively similar if we also include different distance variables in the specifications.
and participation rights are more common if they receive financing from a lead VC located in
another state, unless that state is California. Taken together our results are consistent with a
geographical distance effect that can be traced back to soft information, but also with the
California effect we document.
6. Discussion and Conclusions
This paper shows that geographical elements and regional culture can play an essential
role in contract design in addition the roles played by more “traditional” determinants such as
information and agency problems between contracting parties and legal or other formal
institutions. The VCs we study are sophisticated investors and yet culture and geography seem to
significantly affect their contracting styles, even after we control for founder, company, round,
and VC characteristics. Importantly, unlike international studies of geographical differences in
VC contracts (see Lerner & Schoar, 2005; Kaplan, Martel & Stromberg, 2007; and Bottazi,
DaRin, & Hellmann, 2008), this paper focuses on companies and investors that are located in the
U.S. Therefore, our results cannot be attributed to differences in legal systems, rule-of-law,
accounting transparency, bankruptcy procedures, taxation, etc.
The results presented in this paper can be summarized using a simple hypothetical
example. Consider two software companies. Each one signs a financial contract that accompanies
a VC investment. The first company is headquartered in Silicon Valley and has received
financing from a nearby Silicon Valley VC, whereas the second company is headquartered in
Cleveland and has received financing from a VC operating out of Boston. Suppose that, with the
exception of geographical locations, the observable characteristics of the company, the
entrepreneur and the VC firm are identical. Also, since both companies operate in the U.S., there
are no state-level laws, tax codes, or bankruptcy procedures that affect how contracts have to be
structured. In this example, will the financial contracts for these two companies be different in
practice? Most financial contracting models would argue that the answer should be no, since
geographical factors alone cannot motivate differences in contract design.
The evidence presented in this paper, however, strongly suggests that the answer to this
question is yes. In fact, our analysis shows that the contract in Silicon Valley is likely to be much
less investor-friendly for at least three reasons. First, contract design is affected by the
concentration of the VC market in which the company operates. California is the home of a large
number of VCs and venture-backed companies, and our results show that such a higher
concentration is associated with less investor-friendly contracts. The second factor is the regional
culture and customs of California and Silicon Valley, which we have discussed extensively in the
paper. The third factor is the shorter distance between lender and borrower, which facilitates soft
information and lowers monitoring costs. This latter finding is also consistent with studies that
show that local banks can better serve small businesses.
Appendix A. Detailed Description of Cash Flow Contingencies in VC Contracts
Below is a detailed description of the contract terms we code.
Cumulative Dividends
When a cumulative dividends provision is in force, the VC accrues dividends every year
until the company in which it has invested is sold or liquidated. Cumulative dividends accumulate
and are not paid out in cash to the VC until the company has a liquidation event.19 The dividends
are expressed as a percentage and are typically compounding, which means that investors also
earn dividends on accumulated unpaid dividends. Cumulative dividends are senior to common
stock, and the seniority to other classes of preferred stock is specified in the contract. To illustrate
how cumulative dividends work, consider the following example: Suppose that a VC invests $2
million and receives 8% in compounding cumulative dividends. If the company is sold after five
years for $10 million, then the VC receives (1.085 – 1) × $2 million = $0.94 million in dividends.
As shown in Table 2, 66% of all contracts include no cumulative dividends
(harshness=0). When cumulative dividends are included (harshness=1), the most common
dividend rate is 8%. Our statistics are similar to those found in the Kaplan and Stromberg (2003)
sample, where 44% of all financing rounds have cumulative dividends and the median dividend
rate is the same as in our paper, 8%.
Liquidation Preference
Liquidation preference is the multiple of the investment amount that a VC receives when
the company suffers a liquidation event. Liquidation preference is senior to common stock, and
the seniority to other classes of preferred stock is specified in the contract. Thus, for an
19 A liquidation event could be a merger, acquisition, bankruptcy or other dissolution of the company. Almost all VC contracts include “auto-conversion rights” under which, if the company goes public, an automatic conversion of the VC’s preferred stock to common stock takes place (thus annulling all special contract terms).
investment of $2 million, a liquidation preference of 2X means that the VC gets 2 × $2 million =
$4 million in liquidation preference. Unlike cumulative dividends, the amount that the VC
receives in liquidation preference does not increase over the time.
The majority of all contracts, 93%, have a 1X liquidation preference (harshness=0) and
only 7% have one that is above 1X. The effects of liquidation preference are not specifically
reported by Kaplan and Stromberg (2003).
Participation
All VCs in our sample receive convertible preferred stock. If the preferred stock is not
participating, the VC effectively holds convertible preferred stock and consequently has the
option, at the time of the liquidation event, of receiving either the liquidation preference or
converting the preferred stock to common stock. The fraction of common stock that the VC
receives is determined by dividing the VC’s investment amount by the post-money valuation of
the round. To illustrate how (non-participating) convertible preferred stock works, suppose a VC
invests $2 million at $4 million post-money valuation with a 1X liquidation preference. When the
company is sold, the VC can claim either $2 million in liquidation preference or 50% (2/4) of the
common stock. The VC would choose to convert if and only if the proceeds from the company
were above $4 million.
If the preferred stock is participating, the VC does not have to choose between the
liquidation preference and converting the preferred stock to common stock but instead receives
both. Building on the example, participating preferred stock would give the VC both $2 million
and 50% of the common equity. If the company is sold for $3 million then the VC receives $2
million in liquidation preference and $1 million in common stock (50% of the remaining $2
million).
Participation could either be unconditional, as described above, or conditional on the
amount of the VC cash flows. If the participating preferred stock is “capped” the VC always gets
the common stock but receives the liquidation preference only if the VC’s cash flows are below a
specified multiple or return hurdle, calculated with the VC’s investment as base. To illustrate the
effects of capped participation, suppose that the participation is capped at a 3X gross investment
multiple. If the company is sold for $4 million the VC would receive, with participation, $3
million. Because the gross multiple is 1.5 (3/2), the VC also gets the liquidation preference. If,
however, the company is sold for $18 million, the VC would receive, with participation, $2
million in liquidation preference and $8 million in common stock (50% of $16 million), i.e., a
total of $10 million. Because this would correspond to a gross return of 5X (10/2), which is above
the specified 3X, the VC would not receive the liquidation preference. The total cash flows to the
VC would instead be $9 million (50% of $18 million).
In our sample, 32% of all contracts have (non-participating) convertible preferred stock
(harshness=0) and 68% have either capped or uncapped participating preferred stock
(harshness=1). Participation is less common in the Kaplan and Stromberg sample, with 39% of all
contracts having capped or uncapped participating preferred stock.
Anti-Dilution
If anti-dilution is included in the contract, the VC is issued more preferred stock if and
only if the share price of a follow-up financing round is below the share price that the VC paid in
the earlier financing round. Hence, anti-dilution comes into effect only when the company raises
a follow-up round at a lower valuation. Anti-dilution comes in two forms, weighted average and
full ratchet. Compared with weighted average anti-dilution, full ratchet is more generous to the
VC by issuing more preferred stock, especially if the new financing round is small relative to the
previous round.
Anti-dilution seems to be almost a boiler-plate provision in VC contracts with only 2% of
all contracts having no anti-dilution (harshness=0). Weighted average is most common and found
in 89% of all contracts (harshness=0), while only 9% of contracts have full ratchet anti-dilution
(harshness=1). The Kaplan and Stromberg sample exhibits a somewhat wider distribution of anti-
dilution with 5% of contracts having no anti-dilution, 73% weighted average and 21% full
ratchet.
Redemption
Redemption gives the VC the right to sell back his preferred stock to the company after a
specified number of years. The redemption follows a specified schedule by which, for example,
1/3 of the stock is sold 5 years after the investment, 1/3 after 6 years and the remaining 1/3 after 7
years. In practice, the redemption option is exercised by the VC only if the company is not close
to being acquired or going public. In this situation the company is unlikely to repay the VC the
investment amount so redemption effectively forces the company into bankruptcy.
Redemption is not included in 42% of the sample contracts (harshness=0) while it is
included in 58% (harshness=1). Redemption is more common in the Kaplan and Stromberg
sample and found in 79% of the contracts they study.
Pay-To-Play
The final contract term that we code is pay-to-play, which, unlike the other terms is not
favorable to the VC. When pay-to-play is included in the contract, a VC that chooses not to invest
in follow-up financing rounds of the company is forced to give up some or all of the control and
cash flow contingencies that are attached to the preferred stock. Thus, pay-to-play matters only
when the VC does not invest in a follow-up round.
Pay-to-play is not included in 68% of the sample contracts. Because the VC benefits from
not including pay-to-play in the contract, these contracts are coded as most “harsh”
(harshness=1). Pay-to-play either involves the VC’s losing some contractual rights, typically anti-
dilution, or all contractual rights, forcing her to convert to common stock. Pay-to-play is not
reported by Kaplan and Stromberg (2003).
References
Allen, F. and D. Gale, 1988, Optimal Security Design, Review of Financial Studies 1, 229–263.
Banerjee, A.V. and E. Duflo, 2000, Reputation Effects and the Limits of Contracting: A Study of the Indian Software Industry, Quarterly Journal of Economics 115, 989–1017.
Bengtsson, O., 2008, Relational Venture Capital Financing of Serial Founders, working paper, Cornell University.
Bengtsson, O. and Berk Sensoy, 2008, Investor Abilities and Financial Contracting: Evidence from Venture Capital, working paper, Cornell University and University of Southern California.
Berger, A., N.M. Miller, M.A. Petersen, R. Rajan, and J. Stein, 2005, Does Function Follow Organizational Form? Evidence from the Lending Practices of Large and Small Banks, Journal of Financial Economics 76, 237–269.
Berglof, E., 1994, A Control Theory of Venture Capital Finance, Journal of Law, Economics, and Organization 10, 247–267.
Boot, A. and A. Thakor, 1993, Security Design, Journal of Finance 48, 1349–1378.
Bottazi, L., DaRin, M. and T. Hellmann, 2008b, What is the Role of Legal Systems in Financial Intermediation? Theory and Evidence, Journal of Financial Intermediation, forthcoming.
Casamatta, C., 2003, Financing and Advising: Optimal Financial Contracts with Venture Capitalists, Journal of Finance 58, 2059–2086.
Chisholm, D.C., 1997, Profit Sharing Vs. Fixed Payment Contracts—Evidence from the Motion Pictures Industry, Journal of Law, Economics and Organization 13, 169–201.
Coval, J., and T. Moskowitz, 2001, Geography of Investment: Informed Trading and Asset Prices, Journal of Political Economy 109, 811–841.
Cornelli F., & O. Yosha, 2003, Stage Financing and the Role of Convertible Securities, Review of Economic Studies 70, 1–32.
Cumming, D., 2008, Contracts and Exits in Venture Capital Finance, Review of Financial Studies 21, 1947–1982.
Cumming, D., D. Schmidt and U. Walz, 2008, Legality and Venture Governance Around the World, Journal of Business Venturing, forthcoming.
Degryse, H. and S. Ongena, 2005, Distance, Lending Relationships, and Competition, Journal of Finance 60, 231–266.
Falls, M., 2008, Analyzing VC Deal Terms. Leading Lawyers on Structuring Term Sheets, Developing Negotiation Strategies, and Assessing Risks (United States of America: Aspatore Books).
Fluck, Z., 1998, An Optimal Financial Contracting: Debt vs. Outside Equity, Review of Financial Studies 11, 383–419.
Gilson, R. and D. Schizer, 2002, Understanding Venture Capital Structure: A Tax Explanation for Convertible Preferred Stock, working paper, Columbia University.
Goetzmann, W., S.A. Ravid, R. Sverdlove and V. Pons, 2008, The Role of Soft and Hard Information in the Pricing of Assets and Contract Design—Evidence from Screenplay Sales, working paper, Yale University.
Gompers, Paul, 1998, An Examination of Convertible Securities in Venture Capital Investments, working paper, Harvard Business School.
Gompers, P., J. Lerner, and D. Scharfstein, 2005, Entrepreneurial Spawning: Public Corporations and the Genesis of New Ventures, 1986–1999, Journal of Finance 60, 577–614.
Grinblatt, M. and M. Keloharju, 2001, Distance, Language, and Culture Bias: The Role of Investor Sophistication, Journal of Finance 56, 1053–1073.
Harris, M. and A. Raviv, 1989, The Design of Securities, Journal of Financial Economics 24, 255–287.
Harris, M. and A. Raviv, 1995, The Role of Games in Security Design, Review of Financial Studies 8, 327–3627.
Hart, Oliver, 2001, Financial Contracting, Journal of Economic Literature 39, 1079–1100.
Hellmann, T, 1998, The Allocation of Control Rights in Venture Capital Contracts, RAND Journal of Economics 29, 57–67.
Hellmann, T., 2006, IPOs, Acquisitions, and the Use of Convertible Securities in Venture Capital, Journal of Financial Economics 81, 649–679.
Hochberg, Yael, Alexander Ljungqvist, and Yang Lu, 2007, Venture Capital Networks and Investment Performance, Journal of Finance 62, 251–301.
Hochberg, Yael, Alexander Ljungqvist, and Yang Lu, 2008, Networking as a Barrier to Entry and the Competitive Supply of Venture Capital, working paper, Northwestern University and New York University.
Huberman, G., 2001, Familiarity Breeds Investment, Review of Financial Studies 14, 659–680.
Ivkovich, Z., and S. Weisbenner, 2005, Local does as Local is: Information Content of the Geography of Individual Investors’ Common Stock Investments. Journal of Finance 60, 267–306.
Kaplan, S. N and P. Stromberg, 2003, Financial Contracting Theory Meets the Real World: An Empirical Analysis of Venture Capital Contracts, Review of Economic Studies 70, 281–315.
Kaplan, Steven, and Per Stromberg, 2004, Characteristics, Contracts, and Actions: Evidence from Venture Capitalist Analyses, Journal of Finance 59, 2177–2210.
Kaplan, S. F. Martel and P. Stromberg, 2007, How do Legal Differences and Experience Affect Financial Contracts?, Journal of Financial Intermediation 16, 273–311.
Kedia, S. and Zhou, Z., 2007, Local Market Makers and Trading Costs, working paper, Rutgers University.
Kedia, S., Uysal, V.B., , and V. Panchapagesan, 2008, Geography and Acquirer Returns, Journal of Financial Intermediation 17, 256–275.
Landier, A., 2006, Entrepreneurship and the Stigma of Failure, working paper, Stern School of Business.
Lerner, J., 1995, Venture Capitalists and the Oversight of Private Firms, Journal of Finance 50, 301–318.
Lerner, J., and R. Merges, 1998, The Control of Technology Alliances: An Empirical Analysis of the Biotechnology Industry, Journal of Industrial Economics 46, 125–156.
Lerner, J. and A. Schoar, 2005, Does Legal Enforcement Affect Financial Transactions? The Contractual Channel in Private Equity, Quarterly Journal of Economics 120, 223–246.
Lindsey, L., 2008, Blurring Firm Boundaries: The Role of Venture Capital in Strategic Alliances, Journal of Finance 63, 1137–1168.
Madan, D. and B. Soubra, 2006, Design and Marketing of Financial Products, Review of Financial Studies 4, 361–384.
Malloy, C., 2005, The Geography of Equity Analysis, Journal of Finance 60, 719–755.
Palia, D. S.A. Ravid and N. Reisel, 2008, Choosing to Co-Finance - An analysis of Project Specific Alliances in the film industry, Review of Financial Studies 21, 483–511.
Petersen, M.A., 2004, Information: Soft and Hard, working paper, Northwestern University.
Petersen, M. A. and R. Rajan, 2002, Does Distance Still Matter? The Information Revolution in Small Business Lending, Journal of Finance 57, 2533–2570.
Ravid, S.A. and M. Spiegel, 1997, Linear Securities as Optimal Contracts in an Environment with an Infinite Number of Bad Projects, Journal of Financial and Quantitative Analysis 32, 269–286.
Repullo, R. and J Suarez, 2004, Venture Capital Finance: A Security Design Approach, Review of Finance 8, 75–108
Sahlman, W., 1990, The Structure and Governance of Venture Capital Organizations, Journal of Financial Economics 27, 473–521.
Saxenian, A., 1996, Regional Advantage: Culture and Competition in Silicon Valley and Route 128, Boston: Harvard University Press.
Schmidt, K., 2003, Convertible Securities and Venture Capital Finance, Journal of Finance 3, 1139–1166
Stein, J., 2002, Information Production and Capital Allocation: Decentralized vs. Hierarchical Firms Journal of Finance 57, 1891–1921.
Schultz, P., 2003, Who Makes Markets, Journal of Financial Markets 6, 49–72.
Sorensen, Morten, 2007, How Smart is Smart Money? A Two-Sided Matching Model of Venture Capital, Journal of Finance 62, 2725-2762.
Townsend, R., 1979, Optimal contracts and Competitive Markets with Costly State Verification, Journal of Economic Theory 21, 265–293.
Uzzi, B., 1999, Embeddedness in the Making of Financial Capital: How Social Relations and Networks Benefit Firms Seeking Finance, American Sociological Review 64, 481–505.
Zarutskie, R., 2008, The Role of Top Management Team Human Capital in Venture Capital Markets: Evidence from First-Time Funds, Journal of Business Venturing, forthcoming.
Zender, J., 1992, Optimal Financial Instruments, Journal of Finance 46, 1645–1664.
Figure 1 – Contract Harshness by U.S. State (average Aggregate Contract Harshness)
Dark Grey = Harsh VC Contracts (above 3 Aggregate Contract Harshness) Light Grey = Non-Harsh VC Contracts (3 or below Aggregate Contract Harshness) White = State not in sample
Figure 2 – VC Concentration by U.S. State (based on headquarter location)
Dark Grey = High VC Concentration (20 or more active VCs) Light Grey = Low VC Concentration (Less than 20 active VCs) White = State not in sample
Table 1 - Sample Overview
Number of UniqueContracts 1,800Companies 1,498Lead VCs 628
IndustryRetail Industry 570 32%High-Tech Industry 722 40%Life Science Industry 508 28%
Year of Round2005 218 12%2006 670 37%2007 847 47%2008 65 4%
Company Location (Census 9-Region Division)Pacific 690 38% California 621 35% Silicon Valley 228 13%East North Central 70 4%East South Central 10 1%Mid Atlantic 220 12%Mountain 57 3%North East 329 18%South Atlantic 255 14%West North Central 40 2%West South Central 129 7%
The sample comprises venture capital (VC) financing contracts from U.S. companies that receive financing aU.S. lead VC. Each contract is matched by company name and round date with an investment round listed inVentureEconomics. Company and VC locations refer to headquarters. Industry classification is based on the10-level VentureEconomics classification. Retail Industry includes Communications and Media; ConsumerRelated; Industrial and Energy; and Other Products. High-Tech Industry includes Computer Hardware;Computer Software and Services; and Internet Specific. Life Science Industry includes Biotechnology; andMedical and Health.
Table 2A - Summary Statistics
Deal Conditions # of Obs Mean Median St.DevAggregate Contract Harshness (ACH) 1800 2.59 3.00 1.16Total Round Amount ($ million) 1800 10.79 7.00 12.46Round Number 1800 2.80 3.00 1.55Syndicated Round 1800 0.90Pre-Money Valuation ($ million) 894 48.99 28.47 63.43Fraction of Shares of VCs 894 0.22 0.22 0.11
Company and VC LocationCompany in California 1800 0.35VC in California 1800 0.35Company in Massachusetts 1800 0.16VC in Massachusetts 1800 0.19Company in Texas 1800 0.07Number of Other VCs in California 1800 0.98 1.00 1.22Company in Silicon Valley 1800 0.13VC in Silicon Valley 1800 0.24Distance from Silicon Valley (miles) for non-California VC 1176 42.27 47.61 11.76VC California Investment Experience for non-California VC 1176 0.21 0.15 0.19VC California Syndication Experience for non-California VC 1176 0.32 0.30 0.21
Distance Between VC and CompanyVC and Company Within 5 Miles 1800 0.11VC and Company Within 10 Miles 1800 0.21VC and Company Within 50 Miles 1800 0.42VC and Company in Same State 1800 0.49Distance (miles) 1800 701.00 182.00 94.00
Aggregate Size of VC MarketNumber of VCs in State 1800 374 113 421Number of VCs in Region 1800 474 205 442Number of VC-backed companies in Industry X State 1800 177 119 195Amount of VC financing in Industry X State ($ millions) 1800 1780 1090 1680
Company and Founder CharacteristicsCompany Age 1800 4.13 4.00 2.73Serial Founder 1800 0.22Serial Founder with IPO 1800 0.06Serial Founder with Merger 1800 0.08
See table 1 for overview of sample. Aggregate Contract Harshness (ACH) is the sum of contract termsdiscussed in Table 2B and has a range 0-6. Higher ACH means that the contract is more friendly to the VCinvesting in the round, and especially so if company performance is poor. Variables with unreported medianand standard error are dummy variables.
Table 2B - Overview of Contract Terms
Cumulative Dividends
Included = 1 Non Included = 0Number of Contracts 621 1179Fraction of Sample 35% 66%
Liquidation Preference
Above 1X = 1 1X or Below = 0Number of Contracts 126 1674Fraction of Sample 7% 93%
Participation
Included = 1 Not Included = 0Number of Contracts 1224 576Fraction of Sample 68% 32%
Anti-Dilution
Not Included / Full-Ratchet Weighted Average
Number of Contracts 162 1638Fraction of Sample 9% 91%
Redemption
Included = 1 Not Included = 0Number of Contracts 1044 756Fraction of Sample 58% 42%
Pay-To-Play
Not Included = 1 Included = 0Number of Contracts 1224 576Fraction of Sample 68% 32%
The investor has the right to sell his shares back to the company after a specified time (typically 5-8 years).
Pay-to-play provisions specify contractual rights that the investor loses if he does not invest in a follow-upfinancing round of the company (sometimes only anti-dilution, sometimes all rights).
The multiple of the investor's investment that is paid back to the investor when the company is sold orliquidated. Liquidation preference is senior to common stock.
See table 1 for overview of sample. This table describes individual contract terms and reports their frequency.A coding of 1 means that ACH is 1 unit higher than for a coding of 0.
Dividends that the investor earns annually until the company is sold or liquidated. Cumulative means that thedividends are not paid out annually but when the company is sold or liquidated. Cumulative dividends aresenior to common stock.
With participation the investor receives both a liquidation preference and a fraction of common stock when thecompany is sold or liquidated. With no participation the investor chooses between a liquidation preference anda fraction of common stock.
The investor is issued additional shares if the company raises a new financing round at a lower valuation thanwhat the investor paid (down round). Full Ratchet gives the investor more additional shares than WeightedAverage, especially if the new financing round is small.
Table 3A - Univariate Analysis of Aggregate Contract Harshness
Panel A: VC and Company Location Difference Test
Company in California 2.07 Company outside California 2.86 0.79 ***
Company in Silicon Valley 1.92 Company not in Sil. Valley 2.69 0.77 ***
VC in California 2.15 VC not in California 2.83 0.68 ***
VC in Silicon Valley 2.05 VC not in Silicon Valley 2.76 0.71 ***
VC and Company in 1.84 VC and Company not in 2.81 0.97 ***Silicon Valley Silicon Valley
Panel B: Distance Between VC and Company Difference Test
Distance ≤10 Miles 2.49 Distance >10 Miles 2.62 0.12 *
Distance ≤ 50 Miles 2.53 Distance >50 Miles 2.64 0.11 *
Same State 2.46 Different State 2.72 0.26 ***
Same State 2.00 Different State 2.22 0.22 ** if Company in California if Company in California
VC inside California if Company outside California 2.44
Same State 2.87 Different State 3.03 0.16 ** if Company outside California and VC outside California
Panel C: Company, Founder, VC Characteristics Difference Test
Serial Founder with IPO 2.29 No Serial Founder with IPO 2.61 0.32 ***
VC Experience (> median) 2.44 VC Experience (≤ median) 2.74 0.30 ***
Round Amount Above $7M 2.42 Round Amount Below 2.76 0.35 ***or Equal to $7M
Serial Founder with IPO No Serial Founder with IPOVC Experience (> median) 2.22 VC Experience (≤ median) 2.91 0.68 ***Round Amount Above $7M Round Amount Below
or Equal to $7M
See table 1 for sample description. Mean of Aggregate Contract Harshness (ACH), which is the sum ofcontract terms discussed in Table 2B and has a range 0-6. Higher ACH means that the contract is morefriendly to the VC investing in the round, and especially so if company performance is poor. Rank test ofequality of populations. Significance at 10% marked with *, 5% **, and 1% ***.
Table 3B - Univariate Analysis of Individual Deal Terms
Panel A: VC and Company LocationCum. Dividend Liq. Preference Participation Anti-Dilution Redemption Pay-to-Play
i. Company in Silicon Valley 0.06 0.06 0.67 0.05 0.16 0.92ii. Company outside Silicon Valley 0.38 0.07 0.68 0.09 0.64 0.82
Difference ii-i 0.33*** 0.01 0.01 0.04** 0.48*** -0.010***
iii. VC in Silicon Valley 0.11 0.06 0.62 0.05 0.34 0.87iv. VC outside Silicon Valley 0.42 0.07 0.70 0.10 0.66 0.82
Difference iv-iii 0.31*** 0.02 0.07*** 0.05 0.31*** -0.05*
v. VC and Comp. in Silicon Valley 0.03 0.06 0.65 0.04 0.13 0.92vi. VC and Comp. outside S. Valley 0.44 0.07 0.70 0.10 0.69 0.81
Difference vi-v 0.41*** 0.01 0.04 0.06** 0.56*** -0.11***
Panel B: Distance Between VC and Company
Company in Californiai. VC in Same State 0.07 0.06 0.62 0.05 0.30 0.89ii. VC in Different State 0.18 0.07 0.69 0.09 0.37 0.83
Difference ii-i 0.10*** 0.01 0.06 0.04** 0.07* -0.06**
Company Outside Californiaiii. VC in Different State (non-CA) 0.46 0.05 0.66 0.12 0.75 0.83iv. Same State 0.55 0.08 0.74 0.09 0.75 0.80v. VC inside California 0.28 0.07 0.68 0.07 0.57 0.80
Difference iv-iii 0.09*** 0.03 0.08** -0.03 0.00 -0.02Difference iii-v -0.18*** 0.02 0.02* -0.05 -0.18*** -0.03
See table 1 for sample description. Contract terms are described in Table 2B. Higher variable values means that the contract is more friendly to theVC investing in the round, and especially so on if company performance is poor. Rank test of equality of populations. Significance at 10% markedwith *, 5% **, and 1% ***.
Table 4 - Regression Analysis of VC/Company Location on Aggregate Contract Harshness
Specification 1 2 3 4 5 6 7 8 9 10 11Dependent Variable: ACH ACH ACH ACH ACH ACH ACH ACH ACH ACH ACH
Company in California -1.292*** -0.996*** -1.167*** -1.134*** -0.901*** -1.064*** 0.339 -0.709*** -0.872***[0.106] [0.117] [0.169] [0.127] [0.126] [0.159] [0.932] [0.189] [0.167]
VC in California -0.620*** -0.781*** -0.696*** -0.608*** -0.485** -0.254[0.112] [0.145] [0.121] [0.111] [0.193] [0.241]
Company in Massachusetts -0.257[0.160]
VC and Company in California 0.36[0.224]
VC in Massachusetts -0.201[0.152]
Company in Texas -0.635***[0.199]
Number of Other VCs in California -0.134**[0.055]
Company in Silicon Valley -0.553*** -0.497***[0.178] [0.179]
VC in Silicon Valley -0.353[0.215]
Distance from Silicon Valley (miles) 0.672*** 0.742***[0.121] [0.137]
Distance from S V X Company in California -0.378[0.258]
VC California Investment Experience -1.383***[0.350]
VC California Syndication Experience -1.181***[0.309]
See table 1 for sample description. Ordered logit regressions where the dependent variable is Aggregate Contract Harshness (ACH), which is the sumof contract terms discussed in Table 2B and has a range 0-6. Higher ACH means that the contract is more friendly to the VC investing in the round,and especially so if company performance is poor. Sample in specifications 6-7 includes only companies in California, and in specifications 8-11 onlyVCs in California. Residuals are clustered by company. Significance at 10% marked with *, 5% **, and 1% ***.
Table 4 continued
Specification 1 2 3 4 5 6 7 8 9 10 11
Company Age 0.442*** 0.440*** 0.437*** 0.418*** 0.434*** 0.418** 0.409* 0.450*** 0.452*** 0.434*** 0.412***[0.098] [0.097] [0.097] [0.096] [0.097] [0.210] [0.210] [0.113] [0.113] [0.114] [0.114]
Round Number 0.01 0.011 0.013 0.012 0.017 0.052 0.054 0.007 0.008 0.008 0.02[0.041] [0.041] [0.041] [0.040] [0.041] [0.075] [0.075] [0.050] [0.050] [0.050] [0.050]
Serial Founder -0.14 -0.147 -0.143 -0.11 -0.147 -0.157 -0.195 -0.047 -0.045 -0.035 -0.039[0.161] [0.158] [0.158] [0.157] [0.159] [0.280] [0.277] [0.189] [0.190] [0.187] [0.187]
Serial Founder with IPO -0.078 -0.15 -0.148 -0.156 -0.162 -0.162 -0.128 -0.531* -0.520* -0.510* -0.495*[0.227] [0.233] [0.232] [0.231] [0.231] [0.336] [0.333] [0.275] [0.276] [0.271] [0.272]
Serial Founder with Merger 0.129 0.135 0.139 0.155 0.14 0.221 0.263 0.144 0.151 0.157 0.19[0.195] [0.191] [0.190] [0.190] [0.192] [0.301] [0.296] [0.238] [0.238] [0.238] [0.238]
Number of VCs in Round 0.013 0.015 0.015 0.018 0.036 0.001 0.001 -0.009 -0.009 -0.013 -0.007[0.021] [0.022] [0.022] [0.022] [0.024] [0.035] [0.035] [0.027] [0.027] [0.027] [0.027]
Total Round Amount ($ million) -0.302*** -0.298*** -0.290*** -0.281*** -0.284*** -0.152 -0.143 -0.383*** -0.387*** -0.300*** -0.281***[0.063] [0.063] [0.063] [0.063] [0.064] [0.104] [0.104] [0.080] [0.080] [0.080] [0.081]
VC Number of Investments -0.200*** -0.182*** -0.179*** -0.161*** -0.183*** -0.142** -0.114* -0.211*** -0.208*** -0.164*** -0.167***[0.035] [0.035] [0.035] [0.036] [0.036] [0.056] [0.059] [0.044] [0.044] [0.044] [0.045]
VC Partnership 0.061 0.054 0.046 0.06 0.052 -0.078 -0.062 0.185 0.175 0.153 0.136[0.120] [0.120] [0.120] [0.121] [0.120] [0.222] [0.222] [0.148] [0.148] [0.146] [0.148]
Observations 1800 1800 1800 1800 1800 621 621 1176 1176 1176 1176Sample Full Full Full Full FullPseudo R-squared 0.06 0.07 0.07 0.07 0.07 0.04 0.04 0.05 0.05 0.05 0.05Year and Industry Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Company California VC not in California
Table 5A - Regression Analysis of VC/Company Location on Separate Contract Terms and Other Deal Characteristics
Specification 1 2 3 4 5 6 7 8 9 10 11
Dependent Variable: Dividend Liq.Pref Particip. Anti-Dil Redemp. P-T-P Amount # of VCs Syndic. Valuat. Stake
Company Age 0.157 0.842*** 0.156 0.622*** 0.109 0.454*** 0.124*** -0.038 -0.029 0.268*** -0.028***[0.114] [0.286] [0.109] [0.216] [0.109] [0.144] [0.040] [0.084] [0.141] [0.069] [0.007]
Round Number -0.01 0.189** 0.082* 0.095 0.045 -0.262*** 0.091*** 0.563*** 0.380*** 0.250*** -0.021***[0.050] [0.090] [0.048] [0.076] [0.049] [0.062] [0.017] [0.037] [0.072] [0.028] [0.003]
Serial Founder -0.21 0.327 -0.164 -0.417 -0.013 0.336 0.06 0.115 0.109 0.053 -0.016[0.209] [0.298] [0.194] [0.297] [0.196] [0.258] [0.074] [0.161] [0.295] [0.102] [0.011]
Serial Founder with IPO -0.659** -0.035 -0.126 0.756* -0.098 -0.523 0.235* 0.261 0.573 0.337** -0.006[0.327] [0.452] [0.283] [0.447] [0.294] [0.355] [0.120] [0.271] [0.541] [0.157] [0.015]
Serial Founder with Merger -0.002 -0.323 0.415 -0.001 0.19 -0.318 0.215** 0.347* 0.338 0.281** 0.014[0.280] [0.449] [0.276] [0.462] [0.268] [0.341] [0.100] [0.205] [0.459] [0.137] [0.014]
VC Number of Investments -0.125*** -0.204** -0.064 -0.024 -0.036 -0.166*** 0.071*** -0.012 -0.056 0.092*** 0.002[0.044] [0.081] [0.043] [0.069] [0.044] [0.058] [0.016] [0.034] [0.062] [0.024] [0.003]
VC Partnership 0.082 -0.078 -0.057 0.142 -0.033 0.142 0.06 -0.108 -0.223 0.041 -0.011[0.154] [0.250] [0.145] [0.242] [0.144] [0.185] [0.058] [0.113] [0.230] [0.080] [0.008]
Company in Silicon Valley -1.849*** -0.167 0.1 -0.359 -1.976*** 0.740*** 0.112 -0.108 0.740** -0.148 0.030***[0.336] [0.338] [0.189] [0.374] [0.224] [0.285] [0.072] [0.126] [0.317] [0.097] [0.009]
VC in Silicon Valley -1.391*** -0.066 -0.322** -0.559** -0.967*** 0.388** 0.065 0.033 -0.03 0.166** -0.015*[0.196] [0.259] [0.144] [0.283] [0.143] [0.191] [0.051] [0.110] [0.216] [0.079] [0.008]
Observations 1800 1800 1800 1800 1800 1800 1800 1800 1800 894 894Sample Full Full Full Full Full Full Full Full FullPseudo R-squared 0.12 0.09 0.03 0.06 0.12 0.08 0.13 0.06 0.07 0.31Year and Industry Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes YesRound Amount, Number of VCs Yes Yes Yes Yes Yes Yes No No No No No
See table 1 for sample description. Specifications 1-6 are logit regressions where the dependent variables are separate deal terms (see Appendix A fordescription) that take the value 1 if present and 0 if not present, specification 7 is an OLS regression where the logged total dollar amount of the round is thedependent variable, specification 8 is an ordered logit regression where the dependent variable is the number of VCs in the round, specification 9 is a logitregression where the dependent variable takes the value 1 if the round was syndicated (and 0 otherwise), specification 10 is an OLS regression where thedependent variable is the logged pre-money valuation of the round, and specification 11 is a tobit regression where the dependent variable it the total stakegiven VCs in the round. Residuals are clustered by company. Significance at 10% marked with *, 5% **, and 1% ***. Sample in specification 10-11 includesonly rounds where valuation data is disclosed.
Valuation Data
Table 5B - Regression Analysis of VC/Company Location on Board Seats and Covenants
Specification 1 2 3 4 5 6 7 8 9 10 11
Preferred Common Preferred Number Asset Hire Change Take on Incur Joint RecapBoard Board Board of Sale CEO Business New Debt CapEx Venture Or Reorg
Dependent Variable: Seats Seats Majority Covenant Covenant Covenant Covenant Covenant Covenant Covenant Covenant
Company Age 0.046 0.581*** -0.594 0.238 0.253 0.375 0.18 -0.077 0.18 -0.165 0.09[0.186] [0.206] [0.451] [0.218] [0.188] [0.365] [0.196] [0.181] [0.300] [0.287] [0.236]
Serial Founder -0.782** -0.004 -0.701 0.281 -0.305 -0.152 0.672* 0.128 0.138 0.314 -0.373[0.354] [0.371] [0.839] [0.426] [0.377] [0.737] [0.373] [0.356] [0.543] [0.519] [0.505]
Serial Founder with IPO 1.276** -0.497 -0.43 0.001 1.250** 1.25 -0.482 0.041 -0.582 -0.809 0.952[0.617] [0.681] [1.442] [0.709] [0.605] [1.060] [0.652] [0.595] [1.203] [1.184] [0.729]
Serial Founder with Merger 0.42 0.339 -0.221 0.515 1.036 -0.358 -0.544 0.19 0.355 1.031[0.650] [0.704] [1.426] [0.763] [0.643] [1.232] [0.767] [0.635] [1.247] [0.814]
VC Number of Investments -0.126 -0.062 0.22 0.066 -0.034 0.234 0.068 0.065 -0.355** -0.028 0.203[0.096] [0.102] [0.228] [0.112] [0.096] [0.189] [0.103] [0.094] [0.167] [0.149] [0.128]
VC Partnership -0.191 0.439 0.693 0.469 0.554 0.931 -0.701** -0.219 0.252 0.446 0.08[0.351] [0.348] [1.040] [0.396] [0.356] [0.811] [0.344] [0.334] [0.596] [0.585] [0.481]
Company in Silicon Valley -0.087 -0.16 0.975 -0.396 -0.204 -0.931 0.268 -0.109 -0.298 -0.106 -0.62[0.444] [0.432] [1.162] [0.522] [0.448] [0.923] [0.501] [0.442] [1.201] [0.887] [0.752]
VC in Silicon Valley -0.887** 0.307 -3.570*** -1.224*** 0.234 0.554 -0.845** -1.297*** -1.937* -1.524* -1.468***[0.353] [0.352] [1.288] [0.412] [0.352] [0.609] [0.403] [0.351] [1.151] [0.851] [0.541]
Observations 285 251 141 334 334 334 334 334 334 334 334SamplePseudo R-squared 0.16 0.05 0.26 0.11 0.05 0.08 0.10 0.06 0.15 0.07 0.16Unconditional mean 2.03 1.61 0.20 4.39 0.37 0.07 0.32 0.55 0.10 0.11 0.21Year and Industry Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes YesRound Amount, Number of VCs Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Subsample that includes 344 contracts from first VC financing rounds. Specification 1 is an ordered logit regressions where the dependent variables isthe number of preferred board seats. Specification 2 is an ordered logit regressions where the dependent variables is the number of common boardseats. Specification 3 is a logit regression where the dependent variables takes the value 1 if preferred shareholders hold a majority of the board seats,and 0 otherwise. Specification 4 is an OLS regression where the number of covenants in the contract is the dependent variable (max number ofcovenants is 18). Specification 5-11 are logit regressions where the dependent variables takes the value 1 if the contract includes a covenant and 0 if thecontract does not include the covenant. Residuals are clustered by company. Significance at 10% marked with *, 5% **, and 1% ***.
Subsample with Coded Control Rights - All Contracts are First Round Contracts
Table 6 - Regression Analysis of VC Concentration on Aggregate Contract Harshness
Specification 1 2 3 4 5 6 7 8
Dependent Variable: ACH ACH ACH ACH ACH ACH ACH ACH
Company in California -0.629*** -0.692*** -0.727*** -0.639*** -0.765***[0.179] [0.213] [0.174] [0.167] [0.146]
VC in California -0.618*** -0.664*** -0.610*** -0.627*** -0.628***[0.112] [0.120] [0.112] [0.113] [0.112]
VC in Massachusetts -0.167[0.152]
Company in Massachusetts -0.032[0.177]
Number of VCs in State -0.284*** -0.284*** -0.122*** -0.115*** -0.102**[0.025] [0.025] [0.043] [0.039] [0.044]
Number of VCs in Region -0.128**[0.056]
Number of VC-backed companies -0.167*** in Industry X State [0.053]Amount of VC financing -0.108*** in Industry X State [0.037]
Observations 1800 1800 1800 1800 1800 1800 1800 1800Sample Full Full Full Full Full Full Full FullPseudo R-squared 0.05 0.05 0.07 0.07 0.07 0.07 0.07 0.07Year and Industry Controls Yes Yes Yes Yes Yes Yes Yes YesCompany, Founder, VC Variables Yes Yes Yes Yes Yes Yes Yes YesRegion Controls No No Yes No No No No No
See table 1 for sample description. Ordered logit regressions where the dependent variable is Aggregate ContractHarshness (ACH), which is the sum of contract terms discussed in Table 2B and has a range 0-6. Higher ACH meansthat the contract is more friendly to the VC investing in the round, and especially so if company performance is poor.Residuals are clustered by company. Significance at 10% marked with *, 5% **, and 1% ***.
Table 7 - Regression Analysis of VC/Company Distance on Aggregate Contract Harshness
Specification 1 2 3 4 5 6 7 8 9 10
Dependent Variable: ACH ACH ACH ACH ACH ACH ACH ACH ACH ACH
VC in California -0.670*** -0.760*** -0.702*** -0.783*** -0.487**[0.144] [0.150] [0.157] [0.158] [0.219]
VC and Company Within 5 Miles -0.334 -0.148[0.252] [0.158]
VC and Company Within 10 Miles -0.344* -0.039 -0.390*** -0.504***[0.201] [0.221] [0.133] [0.147]
VC and Company Within 50 Miles -0.650*** -0.111[0.179] [0.123]
VC and Company in Same State -0.588*** -0.283**[0.194] [0.124]
Distance (miles) 0.598*** -0.368*[0.201] [0.206]
Observations 621 621 621 621 621 1179 1179 1179 1179 1179SamplePseudo R-squared 0.03 0.03 0.04 0.03 0.04 0.03 0.04 0.03 0.03 0.04Year and Industry Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes YesCompany, Founder, VC Variables Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
See table 1 for sample description. Ordered logit regressions where the dependent variable is Aggregate Contract Harshness (ACH), which is thesum of contract terms discussed in Table 2B and has a range 0-6. Higher ACH means that the contract is more friendly to the VC investing in theround, and especially so if company exit valuation is low. Residuals are clustered by company. Significance at 10% marked with *, 5% **, and1% ***. Sample in specification 1-5 includes only company in California, and in specifications 6-10 includes only company in other states.
Company California Company non-California